AI applicationsTools and integrationsMem0

Mem0

Mem0 is a framework for managing memory in AI applications. Its graph memory can use Memgraph as the backend, so an agent’s memories are stored and searched as a graph.

Mem0 graph memory with Memgraph works only in Mem0 1.x. Mem0 2.0 removed graph memory from the open-source package, so install a 1.x version, as shown below.

Installation

Install Mem0 1.x with graph memory support:

pip install "mem0ai[graph]<2"

Start Memgraph. Mem0 stores embeddings in a Memgraph vector index:

docker run -p 7687:7687 memgraph/memgraph-mage:latest

Configuration

Set your OpenAI API key:

from mem0 import Memory
import os
 
os.environ["OPENAI_API_KEY"] = "YOUR_OPENAI_API_KEY"

Configure OpenAI for embeddings and Memgraph as the graph store:

config = {
    "embedder": {
        "provider": "openai",
        "config": {
            "model": "text-embedding-3-large",
            "embedding_dims": 1536,
        },
    },
    "graph_store": {
        "provider": "memgraph",
        "config": {
            "url": "bolt://localhost:7687",
            "username": "memgraph",
            "password": "mem0graph",
        },
    },
}

Create the memory store. Mem0 connects to Memgraph and creates the vector index it needs:

m = Memory.from_config(config_dict=config)

Store memories

Add a conversation as memories for a user, with metadata:

messages = [
    {"role": "user", "content": "I'm planning to watch a movie tonight. Any recommendations?"},
    {"role": "assistant", "content": "How about a thriller movie? They can be quite engaging."},
    {"role": "user", "content": "I'm not a big fan of thriller movies but I love sci-fi movies."},
    {"role": "assistant", "content": "Got it! I'll avoid thriller recommendations and suggest sci-fi movies in the future."},
]
 
result = m.add(messages, user_id="alice", metadata={"category": "movie_recommendations"})

Search memories

Search a user’s memories with a question:

for result in m.search("what does alice love?", user_id="alice")["results"]:
    print(result["memory"], result["score"])

The output looks like this:

Loves sci-fi movies 0.315
Planning to watch a movie tonight 0.096
Not a big fan of thriller movies 0.094

For the full example, see Mem0’s Memgraph notebook from the 1.0.5 release.

CogneeClient libraries